9 hours ago
Base Salary
$175k - $296k/yr
Responsibilities
- Design and implement training data pipelines streaming labeled and unlabeled data at hundreds-of-petabytes scale from more than one million vehicles.
- Implement training frameworks for XPENG physical AI foundation models, including VLA 2.0, XWorld, and Robotics.
- Accelerate model training using FSDP, Expert Parallel, Context Parallel, and advanced data types.
- Accelerate cloud model inference for closed-loop simulation, reinforcement learning, and enterprise LLM/VLM applications.
- Profile models and investigate performance bottlenecks to improve training and inference speed.
Requirements
- Master's degree in computer science, computer engineering, or electrical engineering, or equivalent industry experience.
- Deep knowledge of PyTorch.
- Knowledge of model inference frameworks such as vLLM or SGLang.
- In-depth knowledge of transformer architecture and techniques for accelerating transformer training and inference.
- Experience with large-scale distributed model training.
- Preferred experience in the autonomous driving industry.
- Preferred experience writing custom operations with CUDA.
- Preferred experience with edge computing systems.
- Preferred knowledge of distributed computing frameworks such as Ray.
- Ability to solve complex problems collaboratively on larger teams.
Benefits
- Full-time position with a supportive and engaging work environment.
- Access to infrastructure and computational resources to support the work.
- Opportunity to work with leading talent on cutting-edge technology and impact autonomous driving.
- Competitive compensation package with bonus, equity, and benefits.
- Snacks, lunches, dinners, and fun activities.
Tech Stack
PyTorchRobot Framework
Categories
About XPENG
XPENG designs, manufactures, and sells smart electric vehicles for consumers, integrating in-house advanced driver-assistance systems and connected in-car software. Founded in 2014 and headquartered in Guangzhou, it operates plants in Zhaoqing and Guangzhou, maintains a European HQ in Amsterdam, and develops eVTOL aircraft via XPENG AEROHT.